Journal of Environmental Radioactivity 126 (2013) 92e98

Contents lists available at ScienceDirect

Journal of Environmental Radioactivity

journal homepage: www.elsevier.com/locate/jenvrad

Particle size distribution of radioactive aerosols after the Fukushima

and the Chernobyl accidents
Helena Mal*, Petr Rulk, Vera Be
ckov, Jn Mihalk, Miriam Slezkov
National Radiation Protection Institute (SRO), 140 00 Prague, Czech Republic

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 13 January 2013
Received in revised form
12 July 2013
Accepted 28 July 2013
Available online 23 August 2013

Following the Fukushima accident, a series of aerosol samples were taken between 24th March and 13th
April 2011 by cascade impactors in the Czech Republic to obtain the size distribution of 131I, 134Cs, 137Cs,
and 7Be aerosols. All distributions could be considered monomodal. The arithmetic means of the activity
median aerodynamic diameters (AMADs) for articial radionuclides and for 7Be were 0.43 and 0.41 mm
with GDSs 3.6 and 3.0, respectively. The time course of the AMADs of 134Cs, 137Cs and 7Be in the sampled
period showed a slight decrease at a signicance level of 0.05, whereas the AMAD pertaining to 131I
increased at a signicance level of 0.1. Results obtained after the Fukushima accident were compared
with results obtained after the Chernobyl accident. The radionuclides released during the Chernobyl
accident for which we determined the AMAD fell into two categories: refractory radionuclides (140Ba,
140
La 141Ce, 144Ce, 95Zr and 95Nb) and volatile radionuclides (134Cs, 137Cs, 103Ru, 106Ru, 131I, and 132Te). The
AMAD of the refractory radionuclides was approximately 3 times higher than the AMAD of the volatile
radionuclides; nevertheless, the size distributions for volatile radionuclides having a mean AMAD value
of 0.51 mm were very close to the distributions after the Fukushima accident.
2013 Elsevier Ltd. All rights reserved.

Keywords:
AMAD
Aerosol size distribution
Fukushima accident
Chernobyl accident

1. Introduction
The radiological signicance of radionuclides entering the respiratory tract depends on the type, activity and chemical form of
the radionuclide, and also on the aerodynamic properties of the
aerosol particles bearing the radionuclide. These properties can be
expressed by the size distributions of the aerodynamic diameters
(AD). The AD of airborne aerosols lies within a ve-order range,
from 0.001 to 100 mm.
A cascade impactor (CI) can be used to determine the size distribution of the aerosols. CI sorts the particles by their AD, according to the inertial impaction. Detailed information on aerosols,
their behaviour and ways of assessing them can be found in Hinds
(2004).
At the National Radiation Protection Institute (SRO), in Prague,
aerosols and gaseous forms of iodine are continuously sampled by
high volume sampling devices. Moreover, in the event of an
emergency, samples are taken, also by means of CI, so as to determine the size distributions of aerosols-bearing radionuclides.
Samples of this type were taken after the Chernobyl accident and
also after the Fukushima accident.

* Corresponding author. Tel.: 420 226518231.

E-mail address: helena.mala@suro.cz (H. Mal).
0265-931X/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvrad.2013.07.016

This paper presents data on radionuclide activity concentrations

and on the size distribution of aerosols in the air in Prague, Czech
Republic obtained after the Fukushima accident. The results are
compared with data obtained after the Chernobyl accident.
2. Methods e sampling, measurement and evaluation
After the Fukushima accident, samples were taken by means of
three 5-stage cascade impactors (CI), model Sierra Andersen SA
235, on collection substrates (slotted glass bre lters) placed at
each stage and on the glass bre back-up lter situated after the last
stage. The sampling devices were located within the precincts of
SRO in Prague (50
050 N, 14
260 E), about 1.5 m above ground
(concrete surface), in the vertical position, without a cyclone preseparator, protected against rain by a simple shelter. Between 24th
March and 13th April, 5 sets of samples were taken (one set being 3
parallel samples, in one case 2 samples). Each sampling lasted from
3 to 5 days. The collection substrates were changed and the impactors were cleaned within a period of one hour between two
consecutive samplings, and all three parallel CIs were turned on or
turned off within 10 min. The ow rate through the sampling devices was set to 1.88  102 m3/s (40 cfm e cubic feet per minute)
with cut-offs from 0.49 of the 5th stage to 7.2 mm of the 1st stage, in
accordance with the producers recommendation and based on the
calibration provided by the producer. These cut-offs were used in

H. Mal et al. / Journal of Environmental Radioactivity 126 (2013) 92e98

93

Table 1
The activity concentrations of the monitored radionuclides of 5 sampling sets in the monitored period between 24th March and 13th April 2011 after the Fukushima accident
and the meteorological data.
Sampling

Date of sampling

Temperature,
C

No.1
24e27.3
3e19
No.2
27e30.3
1e19
No.3
30.3e3.4
8e24
No.4
3.4e8.4
7e27
No.5
8.4e13.4
6e21
Weighted mean in interval: 24.3e13.4

Intensity of
rainfall, mm/h

Wind speed, m/s

131

0
0
0
0e14.6
0e4.8

0e0.9
0e1.3
0e2.2
0e1.8
0e1.8

1.4 E4
7.7 E4
3.5 E4
2.4 E4
8.8 E5
3.0E-4

the calculations of the AMADs and GSDs (explained below). During

sampling, the ow rate deviated slightly from the set value. However, the readings were frequently tested, the times of the controls
and the ow rates were recorded, and the deviations were
continuously adjusted by a hand-operated ow rate regulator. The
changing ow rate had a slight effect on the cut-off sizes of the ADs
on individual stages. The mean ow rate for each set of parallel
samplings from the recorded data was determined. The cut-offs
calculated for the mean ow rate of an individual set differed by
less than 7% from the cut-offs for a ow rate of 1.88  102 m3/s.
The average deviation was 4% towards the higher cut-off value. The
deviation being only slight, the evaluation was based on the cut-offs
for a ow rate of 1.88  102 m3/s, and the deviation was included
in the uncertainty assessment. The cut-offs are evident from the
tables and gures. After each sampling had been nished, an
ethanol wipe-test was made from the CI stages, around them, and
from the space surrounding the back-up lter, in order to assess the
losses on the impactor walls.
Gamma spectrometry analysis was performed in order to obtain
the activity concentrations deposited on the collection substrate.
We used 5 HPGe detectors of 20e100% relative efciency placed in
shielded cells having either 200 mm steel walls or 100 mm lead
walls, employing the calibration sources for efciency and energy
calibration prepared by the Czech Metrological Institute. The
collection substrates were folded into quarters, and the back-up
lter was folded into sixteenths. For each aerosol size interval,
the folded collection substrates of the parallel samplings of the
respective set were combined into a single measured sample and
measured directly, without any other treatment (close to one
another) on the front part of the detector. The accuracy of the
gamma spectrometry measurements has been regularly checked by
the national metrological authority and the gamma spectrometry
method has been accredited. The spectra were evaluated by the
Canberra GN 2000 program. For comparison, in addition to 131I,
134
Cs and 137Cs natural 7Be was also evaluated. 134Cs activities were
corrected to true coincidences. Although the measurements were
very long (up to 500,000 s), several 134Cs and 137Cs activities were
below the minimum signicant activity (MSA). MSA was determined according to the relation derived by Currie (1968). If no
activity exceeding the MSA level was found for some stage for the
evaluated radionuclide, one half of the MSA level was adopted as
the best activity estimate. This kind of inaccuracy can only slightly
affect the distribution, because the estimated activities were always
very low, of the order of units of percents in relation to the total
activity.
The experimental data e i.e. the radionuclide activity related to
the AD of the aerosol e were evaluated assuming their log-normal
distribution characterized by 2 parameters: activity median aerodynamic diameter (AMAD) and geometric standard deviation
(GSD). AMAD is dened as the value of the aerodynamic diameter
such that 50% of the airborne activity is associated with particles
smaller than AMAD, and 50% of the activity is associated with

I, Bq/m3

134

Cs, Bq/m3

137

Cs, Bq/m3

6.6 E6
6.7 E5
3.0 E5
3.2 E5
1.7 E5
2.9E-4

8.6
7.6
3.3
2.9
1.8
2.9

E6
E5
E5
E5
E5
E5

Be, Bq/m3

2.9
3.0
3.3
2.9
2.0
3.1

E3
E3
E3
E3
E3
E3

particles larger than AMAD. Both parameters were assessed from

the linear regression of cumulative activities (expressed as a
quantile of the normal distribution) on the AD logarithm. The
method is described in detail as the inverted probit method in
OShaughnessy and Raabe (2003).
3. Results and discussion
3.1. Monitoring after the Fukushima accident
The trajectories of the air masses from Fukushima to Europe
are generally described in Masson et al. (2011) and Thakur et al.
(2013). The character of the aerosol size distributions depends,
among other things, on the time that it stays in the atmosphere.

Table 2
The fraction of the activity concentrations of the monitored radionuclides found in
the individual size intervals of ADs after the Fukushima accident (6 back-up lter).

Sampling No.1

Sampling No.2

Sampling No.3

Sampling No.4

Sampling No.5

Arithmetic mean of
fractions weighted
by the time of sampling

Stage

Size
interval
[mm]

131

I
[%]

134

Cs
[%]

137

Cs
[%]

7
Be
[%]

1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
5
6

>7.2
3.0e7.2
1.5e3.0
0.95e1.5
0.49e0.95
<0.49
>7.2
3.0e7.2
1.5e3.0
0.95e1.5
0.49e0.95
<0.49
>7.2
3.0e7.2
1.5e3.0
0.95e1.5
0.49e0.95
<0.49
>7.2
3.0e7.2
1.5e3.0
0.95e1.5
0.49e0.95
<0.49
>7.2
3.0e7.2
1.5e3.0
0.95e1.5
0.49e0.95
<0.49
7.2
3.0e7.2
1.5e3.0
0.95e1.5
0.49e0.95
<0.49

2
1
8
12
26
51
1
3
5
11
27
53
1
4
6
10
22
57
1
5
9
14
18
53
2
6
11
17
15
49
1
4
8
13
20
54

4
4
4
25
38
25
1
4
21
11
20
43
3
6
3
12
26
50
1
3
2
10
19
65
1
1
1
7
25
65
2
3
5
12
25
53

5
3
3
22
28
39
1
3
24
12
19
41
3
7
5
11
27
47
1
5
2
10
19
63
1
3
4
7
25
60
2
4
6
12
23
53

1
3
6
17
34
39
1
2
5
15
31
46
1
3
5
14
30
47
1
3
5
12
20
59
1
2
3
9
23
62
1
2
5
13
27
52

94

H. Mal et al. / Journal of Environmental Radioactivity 126 (2013) 92e98

Leakage of radionuclides from NPP Fukushima started on 12th

March 2011, and the contaminated air mass reached the Czech
Republic on 23rd March 2013, i.e. 11 days later (the direct distance
from Fukushima to Prague is approximately 9000 km).
The activity concentrations of the evaluated nuclides and the
sampling periods are shown in Table 1; the peak values were
reached in set 2. The combined standard uncertainties of the activities of the radionuclides in all measurements ranged from 6 to
30%; they were higher only in exceptional cases. Several activity
values were below MSA. MSA was about 0.5 mBq/m3 for both Cs
isotopes, and 3 mBq/m3 for 131I. Detailed monitoring results for the
activity concentrations of air masses for the territory of Europe are
summarized in Masson et al. (2011).
The fractions of the activity concentrations of the monitored
radionuclides found in the individual size intervals of ADs are
shown in Table 2. The highest activity fraction was observed in the

0.6

131I

Fraction

0.5

0.4

0.3

0.2

0.1

0.0
0.01 - 0.49

0.49 - 0.95

0.95 - 1.5

1.5 - 3.0

3.0 - 7.2

7.2 - 100

AD [m]

0.6

134Cs

1
2

Fraction

0.4

3
4

0.2

5
0.0
0.01 - 0.49

0.49 - 0.95

0.95 - 1.5

1.5 - 3.0

3.0 - 7.2

7.2 - 100

AD [m]

0.6
137Cs

1
2
3
4
5

Fraction

0.4
0.2
0.0
0.01 - 0.49

0.49 - 0.95

0.95 - 1.5

1.5 - 3.0

3.0 - 7.2

7.2 - 100

AD [m]

Fraction

0.6

7Be

1
2

0.4

3
4

0.2

5
0.0
0.01 - 0.49

0.49 - 0.95

0.95 - 1.5

1.5 - 3.0

3.0 - 7.2

Sampling
No.1

0.6

Size distributions of average activity

0.3

I 131
Cs 134
Cs 137

0.2

Be 7

0.5

Fraction

Table 3
AMAD [mm] and GSD values of 5 sampling sets after the Fukushima accident
(AM arithmetic mean from log-normal distribution).

7.2 - 100

AD [m]

smallest particle size interval with AD <0.49 mm. It was usually over
40% e with the exception of 134Cs in the rst set, where the
maximum activity was in the second smallest size interval. The
fraction of the activity of a given nuclide present in the aerosol size
interval with AD <0.49 mm obtained from the weighted activity
average of all 5 sets (the weight having been the sampling period)
for all nuclides exceeded 50%. For aerosols with AD <0.95, the value
was around 75%, and for aerosols with AD >3 mm the value was less
than 10%.
The size distribution of the 137Cs transported from Fukushima
could be affected by 137Cs from the pre-Fukushima period. However, this effect is not signicant. Before the Fukushima accident,
the average 137Cs activity concentration in the air (at the same
sampling site) was usually below 106 Bq/m3 (see www.suro.cz),
whereas the 137Cs activities in the samples taken for determining
the size distribution were higher than 105 Bq/m3.
Fig. 1aee shows the ratio of the activity concentration to the
logarithmic width of the size interval dA/dlog AD normalized towards one related to AD (The logarithmic width of the interval is
given by the difference between the logarithm of the maximum
and minimum of this interval.). This representation is based on the
assumption of a log-normal distribution of activities with respect
to AD (which is often the case with aerosols), and it thus offers a
better picture for assessing the character of the distribution,
including the modality. Due to the low probability of the occurrence of particles below 0.01 mm and above 100 mm, the smallest
size interval was taken from 0.01 mm to 0.49 mm and the largest
interval was from 7.2 mm to 100 mm. The choice of the limiting
values (0.01 mm and 100 mm) does not affect the AMAD and GSD
calculations.
When relating activity to the logarithmic width of the interval,
the maximum activity shifted from the size interval with AD
<0.49 mm towards the size interval of 0.49e0.95 mm and 0.95e
1.5 mm (see Fig. 1). With 131I, there was a slight increase in the
fraction with AD 0.95e7.2 mm with time, and a decrease in the
fraction with AD 0.49e0.95 mm. This means that the distributions
shifted towards larger ADs. With caesium, an increase can be seen
in the two smallest size fractions; it is interesting to note that there
is also an increase in the smallest size fraction for 7Be. Changes in
size distribution in time are discussed below.
The size distributions of the aerosols calculated from the mean
weighted activities over the whole sampling period, from 24th
March until 13th April (Fig. 1e) of all nuclides, including 7Be, are in
fact identical. The distributions can be regarded as monomodal in
the whole range for all radionuclides, even if we are aware that the
particles under 0.49 mm often contain more than 50% of the activity,
and that we do not know anything about the inner distribution

0.4

0.1
0.0
0.01 - 0.49

0.49 - 0.95

0.95 - 1.5

1.5 - 3.0

3.0 - 7.2

7.2 - 100

AD [m]
Fig. 1. aee. The ratio of the activity concentration to the logarithmic width of the size
interval dA/dlog AD normalized towards one related to AD for 5 sampling periods after
the Fukushima accident.

131

AMAD
0.38
GSD
3.6
No.2
AMAD
0.36
GSD
3.3
No.3
AMAD
0.35
GSD
3.8
No.4
AMAD
0.47
GSD
3.4
No.5
AMAD
0.53
GSD
3.4
AM
AMAD
0.42
GSD
3.5
AM of AMAD values for 131I, 134Cs and 137Cs
AM of GSD values for 131I, 134Cs and 137Cs

134

Cs

0.69
3.1
0.66
2.6
0.36
4.6
0.25
3.7
0.17
4.0
0.43
3.6

137

Cs

0.48
4.3
0.71
2.3
0.40
4.2
0.31
3.5
0.27
4.3
0.43
3.7
0.43
3.6

Be

0.53
2.6
0.46
2.7
0.44
3.2
0.34
3.3
0.28
3.3
0.41
3.0

H. Mal et al. / Journal of Environmental Radioactivity 126 (2013) 92e98

95

within the range below 0.49 mm, where the particles could be
distributed in a more complicated manner. Our results are valid,
provided that the distribution is monomodal in the whole range.
Table 3 shows the AMAD and GSD values of the analysed distributions. The AMADs of the articial radionuclides ranged from
0.17 mm to 0.71 mm, and the GSDs ranged from 2.3 to 4.6. The
arithmetic means of the AMAD and the GSD of the analysed radionuclides obtained from the parameters of log-normal distribution of the AMADs and GSDs are practically identical. Close values
were also obtained when assessing the AMADs and GSDs from the
weighted activity means shown in Table 2. The distribution of all
AMAD and GSD values of man-made radionuclides in the quantile
graph for log-normal distribution is given in Fig. 2. The arithmetic
mean was estimated from its parameters, as it can be seen that the
data are log-normally distributed without outliers. The AMAD
arithmetic mean determined from the log-normal distribution
parameters was 0.43 mm, and the GSD mean was 3.6. The AMAD
values for 7Be lay between 0.28 mm and 0.53 mm, the arithmetic
mean being 0.41 mm; for GSD the values lay between 2.6 and 3.3,
the mean being 3.0. The similarity in the size distributions of the
aerosols related to man-made radionuclides and to naturally
occurring 7Be can be due to the fact that an aerosol of this size
(tenths of mm) is retained in the atmosphere for quite a long time,
and is thus easily transported over large distances.
The factors that contribute to the uncertainty assessment of
AMADs are as follows: uncertainty of activity determination, losses
on the inner walls of the CI, inter-stage losses due to rebounds and
re-entrainment of the particles, instability of the air stream, and
uncertainty of the ow rate calibration. The signicance of the
aerosol deposition on the inner walls of CI was estimated by
measuring the activity of wipe-test samples taken from the inner
surface of CI after each sampling. These activities amounted to at
most several percent of the total activity measured on the collection substrates. No time trend in radionuclide deposition on the
inner surfaces of the impactor was observable.
The uncertainty assessment of the AMAD and GSD values was
performed in the following way: an activity simulation for a given
radionuclide in each set was performed for each size interval,
assuming normal activity distribution with parameters N(A, s2),
where A stands for the measured activity and s is its combined
uncertainty in terms of standard error. The generated activity
values for each size fraction were then used to obtain the AMADs
and GSDs in a standard way (described in Methods) from the
linear regression of cumulative activities (expressed as a quantile)
on AD logarithm (OShaughnessy and Raabe, 2003). These simulations were performed 1000 times, so that 1000 AMAD and GSD
values were obtained for each radionuclide in each set. The mean
AMAD and GSD and their standard deviations were estimated from
these 1000 values. The standard deviations were increased by the
uncertainty of the ow rate assessment, using the error expansion
method. Typically, the combined uncertainties of the AMADs were

from 6 to 16%, with a mean of 13%. The combined uncertainty of

GSD was from 5 to 26%, with a mean value of 13%.
The time course of the AMAD and GSD values is presented in
Fig. 3a and b. It was analysed by the weighted linear regression
method. The weight taken into account was the reverse value of the
square of the uncertainty assessment of the given value. The
AMADs of 134Cs, 137Cs, and 7Be in the monitored period were
decreasing at signicance level of 0.05. At this level, the 131I trend
was not signicant, but it did gradually increase at the level of 0.1.
The time course of the GSD values featured a slight increase at a
statistical signicance level of 0.05 only for 137Cs and 7Be. The
gradual decrease in the AMAD of the caesium isotopes and the less
signicant decrease in 7Be bearing aerosols may be due to a quicker
reduction in the occurrence of larger aerosols with time, a phenomenon possibly caused by the combined effect of rain (4th
sampling and 5th sampling e a total amount of about 5 mm and
3 mm of rain, respectively). The increase in AMAD with time, as
found for 131I, could be explained by the reaction of gaseous 131I on
the aerosol surface, on the assumption that a coarse aerosol had
more reaction-prone surfaces than a ner aerosol, and that this
reaction process exceeded the removal process of larger aerosols
from the atmosphere with time, as in the case of Cs isotopes.
To determine the committed effective dose for the population
from inhalation of aerosols, AMAD is mostly taken equal to 1 mm
(ICRP Publication 66, 1994). The committed effective dose from
inhalation obtained with the IMBA program (IMBA Professional
Plus program) for the mean Fukushima value of AMADs of
0.43 mm observed in the Czech Republic is two thirds of the value
for 1 mm.
Until now, there has been a shortage of published aerosol size
distribution results related to the Fukushima accident. Kaneyasu
et al. (2012) presented the results of 2 measurements of the size
distributions of 134Cs and 137Cs in aerosols collected after the accident; in the rst measurements (28th Aprile12th May) Kaneyasu
obtained AMADs 0.54 and 0.53 mm, and in the second measurements (in May) they obtained 0.63 mm for both Cs isotopes.
They found that sulphate was the potential transport medium for

Fig. 2. Distribution of all AMAD and GSD values of articial radionuclides in the
quantile graph for log-normal distribution after the Fukushima accident.

Fig. 3. a,b. Time course of the AMAD and GSD values of

Fukushima accident.

131

I,

134

Cs,

137

Cs after the

96

H. Mal et al. / Journal of Environmental Radioactivity 126 (2013) 92e98

these radionuclides, and resuspended soil particles with attached

radionuclides were not the major airborne radioactive substances
at the time of measurement. Kanai (2011) described how the activity ratio of atmospheric 134Cs/137Cs in aerosols decreased over
time, almost following the physical decays. Koizumi et al. (2012)
presented the size distribution of 134Cs and 137Cs activities in 9
size intervals from samplings taken in the Fukushima region between July 2nd and July 8th, 2011 i.e. more than 3 months after the
accident. The distributions were signicantly bimodal, with the
mode boundary between 1.1 and 2.2 mm. Based on these distributions we assessed the following values: AMAD of smaller mode
0.52 mm and 0.48 mm for 134Cs and 137Cs, and GSD 1.5 and 1.6
respectively; and AMAD of larger mode 5.8 mm and GSD 1.8, for
both isotopes of Cs. The AMAD of the smaller mode is close to the
AMAD observed by our team, and the GSD value is roughly one half.
The AMAD of the larger mode most probably refers to a resuspended aerosol, which has a rather short-term presence in the
atmosphere, and is not transported over long distances.
3.2. Monitoring after the Chernobyl accident
The results of monitoring of the aerosol size distribution after
the Chernobyl accident were published in a brief form in Rulk et al.
(1989); the data have now been processed in a different way, and
the results presented here are more detailed.
The distance between Prague and Chernobyl, as the crow ies, is
approximately 1100 km. The trajectory and the mixing of the air
masses after the Chernobyl accident were quite complex, and 3
main passages of contaminated air masses over the territory of the
Czech Republic and the Slovak Republic (former Czechoslovakia)
were observed. The rst passage, from 30th April 1986 came from
the leak of 27th April 1986; the second maximum passage of 3rd
Maye4th May 1986 came from the leak of 26th April 1986 (i.e. this
leakage took place one day earlier than the leakage for the rst
passage); and the third passage of 7th Maye9th May 1986 came
from the leak of 4th Maye5th May 1986. The activities in the rst
passage were from several Bq/m3 to tens of Bq/m3, depending on
the radionuclide, the activities of the second passage were about
one-third of the activities of the rst passage, and the activities of
the third passage were around one-tenth of the activities in the rst
passage (De Cort et al., 1998). From these data it can be estimated
that, at the time of sampling by CI, the entrapped radionuclides had
been released from Chernobyl for more than 8 days before they
were captured by CI. Detailed information on the trajectories is
given in Report (1986).
Following the Chernobyl accident, 9 samplings were taken between 3rd May and 20th June 1986, making use of the same 5-stage
CI as after the Fukushima accident. The total activity concentrations
of 131I, 134Cs, and 137Cs ranged between about 6 Bq/m3 (3.5 for Cs
and 6.1 Bq/m3 for 131I) in the rst sampling and 104 Bq/m3 in the
last sampling. In addition to the nuclides that were analysed after
the Fukushima accident, other radionuclides were also analysed
after the Chernobyl accident. Radionuclides can be classied into
two categories, i.e. volatile, released into the environment during
the re and then condensing, and refractory, scattered into the
atmosphere by the explosion. The rst group included 134Cs, 137Cs,
103
Ru, 106Ru, 131I and 132Te; the second group included 140Ba, 140La,
141
Ce, 144Ce, 95Zr and 95Nb. The data on 95Nb activity were corrected
for the build-up of 95Nb from 95Zr during the time between the
collection of the sample and the measurement of the sample.
Fig. 4aed shows the ratio of the activity concentration to the logarithmic width of the size interval dA/dlog AD normalized towards
one related to AD for 131I, 134Cs, 137Cs and 95Nb (as a representative
of refractory radionuclides) for the 1st and 2nd sampling. For the
3rde9th sampling, combined activity concentrations were

Fig. 4. aed. The ratio of the activity concentration to the logarithmic width of the size
interval dA/dlog AD normalized towards one related to AD after the Chernobyl accident.

determined as the weighted means of the activity concentrations

from the individual samplings, where the weight was the sampling
period. The 1st and 2nd samples were obtained in smaller towns in
the CR, while samples 3e9 were obtained at the same site in Prague. The activities of 95Nb in samples No. 4, 7 and 9 lay below MSA,
and they were not included in calculations. All distributions were
roughly monomodal, with the exception of volatile radionuclides
from the 1st sampling, which were bimodal; the boundary between
modes lay in the size interval of AD 0.49e0.95 mm.
With the exception of the refractory radionuclides from the
1st and 2nd sampling, the share of aerosol activity was highest in
the AD <0.49 mm size interval. With the volatile radionuclides,
the share of activities in the smallest size particle fraction was
usually higher than 40e50%; in the fraction of particles smaller
than 0.95 mm, it exceeded 80%. The proportion of the activity of
AD >3 mm particles was about 15%. The size distributions of the
refractory radionuclide were shifted towards larger sizes
(Table 4).

H. Mal et al. / Journal of Environmental Radioactivity 126 (2013) 92e98

97

Table 4
AMAD [mm] and GSD values of 9 samplings after the Chernobyl accident (AM arithmetic mean from log-normal distribution).
141

Sampling
No.1
No.2
No.3
No.4
No.5
No.6
No.7
No.8
No.9
AM of 3e9

AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD
AMAD
GSD

Ce

2.00
3.2
1.08
2.7

144

Ce

1.87
3.2
1.08
2.3

134

Cs

0.44
3.4
0.48
3.0
0.49
3.0
0.60
3.3
0.51
4.0
0.96
3.4
0.15
8.7
0.60
5.4
0.24
7.0
0.51
3.9

137

Cs

0.47
3.2
0.49
3.1
0.48
2.8
0.59
3.3
0.52
4.1
1.00
3.4
0.16
8.1
0.65
5.9
0.35
8.8
0.52
3.9

The AMADs of volatile radionuclides ranged between 0.15 and

1.0 mm (the AMADs were calculated from all size intervals also for
radionuclides from the 1st sampling, which were bimodal). The
AMADs of refractory radionuclides were between 0.4 mm and
4.0 mm. The GSD values for all nuclides ranged in a wide interval
from 1.7 to 8.8. All AMAD and GSD distributions were approximately log-normal, though in the case of GSD the conditions for
log-normality to be satised were slightly worse. The arithmetic
means of AMADs obtained from log-normal distribution parameters for volatile radionuclides and refractory radionuclides were
0.51 mm and 1.5 mm, respectively; i.e. the AMAD of the refractory
radionuclides was approximately 3 times greater than the AMAD of
the volatile radionuclides. The arithmetic means of GSD were 4.1
for volatile radionuclides and 3.2 for refractory radionuclides. Disregarding the AMAD and GSD from the 1st sampling of volatile
radionuclides with a clearly bimodal distribution had almost

103

Ru

0.35
4.1
0.58
2.9
0.30
3.7
0.43
4.2
0.40
3.6
0.83
2.8
0.17
5.6
0.80
3.7
0.42
3.8
0.40
3.7

106

Ru

0.37
4.0
0.57
2.9

0.42
4.8
0.91
2.9

0.55
3.0
0.56
3.5
0.39
3.9

131

0.44
5.1
0.25
4.0
0.48
5.0

132

Te

0.54
3.2
0.69
2.5
0.30
4.2

0.47
3.4

0.40
4.8

0.30
4.2

95

Nb

95

Zr

3.05
3.2
1.70
1.8

4.03
3.6
1.59
1.7

1.10
3.9
0.93
4.3
2.13
4.5

1.95
1.9
1.27
5.3

0.34
9.2

1.86
3.8

BaeLa
1.06
2.8
0.81
2.8
0.36
4.0
0.76
3.0
0.51
3.5
1.27
2.9

0.52
3.5

no effect on the means of the AMADs and GSDs of volatile

radionuclides.
The time course of the AMADs and GSDs of 131I, 134Cs, 137Cs and
95
Nb is shown in Fig. 5a and b. The time course of the AMADs was
imperceptible; the GSDs had a slightly growing trend up to the 6th
sampling; then, however, they increased signicantly. The increase
in GSD could be due either to the longer sampling interval with
changes in size distribution, or to higher uncertainties of the activity determinations.
A great deal has been published about the Chernobyl accident.
For example, Reineking et al. (1987) and Bondietti and Brantley
(1988) provided information on the growth rate of the median
diameter of the aerosols released into the atmosphere during their
interaction with the chemical and physical environment of the
lower atmosphere. Tschiersch and Georgi (1987) wrote about differences between 131I and 137Cs; caesium was transported in the
form of particles, but 131I was mainly transported in the gaseous
phase, and was adsorbed to an individual local aerosol throughout
the journey from Chernobyl. The distributions are slightly shifted to
a larger diameter with increasing transport distance. For iodine
alone, there was a much smaller increase in the median. It was
observed (Muck et al., 2002) that the iodine-to-caesium activity
ratio changed with distance (from Chernobyl to Central and
Northern Europe) by almost one order of magnitude. This is
explained by the higher deposition velocity of iodine. Results
summarizing the AMAD determination in the environment can be
found in Dorrian and Bailey (1996) and Dorrian (1997). 78 results
came from monitoring of the Chernobyl accident, with a range of
0.3 mm and 4.8 mm and a median of 0.64 mm; and 16 results came
from resuspended aerosols, which had a range of 1.0 mm and 15 mm
and a median of 6.0 mm. Our measurement results obtained after
the Chernobyl accident are in good agreement with the AMADs and
the median ranges obtained by other authors after Chernobyl.
4. Conclusions

Fig. 5. a,b. Time course of the AMAD and GSD values of 131I, 134Cs, 137Cs and 95Nb after
the Chernobyl accident.

The direct distance from Fukushima to Prague (9000 km) is

about 8 times greater than the distance between Chernobyl and
Prague (1100 km), but the time that the aerosol spent in the atmosphere before the particulate matter was trapped in the lter in
Prague was similar, more than 8 days.
The activity concentrations on the Czech territory after the two
accidents differed by many orders of magnitude.

98

H. Mal et al. / Journal of Environmental Radioactivity 126 (2013) 92e98

After the Fukushima accident we obtained size distributions of

man-made 131I, 134Cs, 137Cs and natural 7Be in 5 sets of samplings.
All distributions can be regarded as monomodal. The arithmetic
means of the AMADs and GSDs of 131I, 134Cs, and 137Cs from all sets
of samplings were 0.43 mm and 3.6, respectively. The mean 7Be
AMAD value was 0.41 mm, and the mean GSD value was 3.0. The
aerosol size distribution of the man-made radionuclides was very
close to that of 7Be. The average combined uncertainty assessment
of AMAD and GSD was 13%.
After the Chernobyl accident, the size distributions included more
radionuclides, namely the group of refractory radionuclides (140Ba,
140
La, 141Ce, 144Ce, 95Zr, and 95Nb) and the group of volatile radionuclides (134Cs, 137Cs, 103Ru, 106Ru, 131I, and 132Te). All distributions
were roughly monomodal, with the exception of the volatile radionuclides from the 1st sampling, which were bimodal; the boundary
between the modes lay in the size interval of AD 0.49e0.95 mm.
The AMADs of refractory radionuclides were approximately
three times higher than those of volatile radionuclides. The volatile
nuclide size distributions with a mean value of AMAD 0.51 mm and
GSD 4.1 were very similar to the distributions obtained after the
Fukushima accident. Due to the scatter of the values, the difference
between the mean AMAD values is insignicant.
In both accidents, the activity of 131I, 134Cs and 137Cs present in
the fraction of the smallest particles having AD <0.49 mm usually
exceeded 40% of their total activity. In both cases, aerosols with AD
<0.95 mm bore 75e80% of the aerosol activity. The activity seen on
aerosols with AD >3 mm after the Fukushima accident was below
10%, and after Chernobyl the level was roughly 15%.
The AMAD values measured after the Chernobyl accident are in
good agreement with the values given in Dorrian and Bailey (1996)
and Dorrian (1997). The AMADs of the smaller mode of the aerosol
bimodal distribution sampled in the Fukushima region (Koizumi
et al., 2012) three months after the accident were 0.48e0.52 mm,
and were thus close to the values measured in the Czech Republic
shortly after the Fukushima accident.
The time course of the AMADs for 134Cs, 137Cs, and also for
naturally occurring 7Be after the Fukushima accident displayed a
slight statistically signicant decrease at the signicance level of
0.05, and the time distribution of the AMADs of 131I increased
gradually at a signicance level 0.1. No time trend for AMADs after
the Chernobyl accident was observed.
The time course of the GSDs following both accidents showed a
slight increase; moreover, towards the end of the monitoring
period after the Chernobyl accident the GSDs increased by leaps.
The increase in GSD with time can be accounted for both by the
increase in activity uncertainty in each fraction as the activity
decreased, and by the longer sampling times, due to which each
sampling could register aerosols with more diverse AD.
Despite the different conditions affecting the formation of
radioactive aerosols during the two accidents, despite the trajectory lengths as well as various meteorological effects, the size distributions of the aerosols associated with 131I, 134Cs and 137Cs
monitored on the territory of the Czech Republic were very similar,
probably due to the monitoring of aerosols of similar age. Aerosols

of this size (tenths of mm) are retained in the atmosphere for quite a
long time, and thus can be transported over long distances. They
were also close to the 7Be size distributions.
Funding
This investigation was supported by the project Research of
advanced methods for detection, assessment and consequential
management of radioactive contamination, identication code VF
20102015014, funded by the Ministry of the Interior of the Czech
Republic.
References
Bondietti, E.A., Brantley, J.N., 1988. Size distribution and growth of natural and
Chernobyl-derived submicron aerosol in Tennessee. J. Environ. Radioact. 6.
Currie, L.A., 1968. Limits for qualitative detection and quantitative determination.
Anal. Chem. 40, 586e593.
De Cort, M., et al., 1998. Atlas of Cesium Deposition on Europe after the Chernobyl
Accident. Ofce for Ofcial Publications of the European Communities, ECSCEEC-EAEC, Brussels-Luxemburg.
Dorrian, M.D., Bailey, M.R., 1996. Particle size distributions of radioactive aerosols
measured in the workplace and the environment. J. Aerosol Sci. 27 (Suppl. 1),
439eS440.
Dorrian, M.D., 1997. Particle size distributions of radioactive aerosols in the environment. Radiat. Prot. Dosim. 69 (2), 117e132.
Hinds, W.C., 2004. Aerosol Technology: Properties, Behavior and Measurement of
Airborne Particles, second ed. J. Wiley-Interscience Publication, John Wiley &
Sons, Inc., ISBN 0-471-19410-7
www.imbaprofessional.com.
International Commission on Radiological Protection, 1994. Human Respiratory
Tract Model for Radiological Protection. In: ICRP Publication, vol. 66. Elsevier
Science Ltd., Oxford. Ann. ICRP 24(1e3).
Kanai, Y., 2011. Monitoring of aerosols in Tsukuba after Fukushima Nuclear Power
Plant incident in 2011. J. Environ. Radioact. 111, 33e37.
Kaneyasu, N., et al., 2012. Sulfate aerosol as a potential transport medium of
radiocesium from the Fukushima nuclear accident. Environ. Sci. Technol. 46,
5720e5726.
Koizumi, A., et al., 2012. Preliminary assessment of ecological exposure of adult
residents in Fukushima Prefecture to radioactive cesium through ingestion and
inhalation. Environ. Health Prev. Med.. http://dx.doi.org/10.1007/s12199-0110251-9.
Masson, O., et al., 2011. Tracking of airborne radionuclides from the damaged
Fukushima Dai-Ichi nuclear reactors by European networks. Environ. Sci.
Technol. 45 (18), 7670e7677. http://dx.doi.org/10.1021/es2017158.
Mck, et al., 2002. A consistent radionuclide vector after the Chernobyl accident.
Health Phys. 82, 141e156.
OShaughnessy, P.T., Raabe, O.G., 2003. A comparison of cascade impactor data
reduction methods. Aerosol Sci. Technol. 37, 187e200. http://dx.doi.org/
10.1080/02786820390112506.
Reineking, A., Beckert, K.H., Porstendrfer, J., 1987. Air activity concentrations and
particle size distributions of the Chernobyl aerosol. Radiat. Prot. Dosim. 19 (3),
159e163.
Report, 1986. 1986 on Radiation Situation in CSSR after Chernobyl Accident, vol. 48.
Institute of Hygiene and Epidemiology, Centre of Radiation Hygiene, Prague,

Srobrova.
Rulk, P., Maltov, I., Bu
cina, I., 1989. Aerosol particle size distribution in dependence on the type of radionuclide after the Chernobyl accident and in the NPP
efuents, the radioecology of natural and articial radionuclides. In: Proceedings of the 15th Regional Congress of IRPA, Visby, Gotland, Sweden. ISSN:
1013-4506, pp. 102e107.
Thakur, P., Ballard, S., Nelson, R., 2013. An overview of Fukushima radionuclides
measured in the northern hemisphere. Sci. Total Environ. 458e460, 577e613.
Tschiersch, J., Georgi, B., 1987. Chernobyl fallout size distribution in urban areas.
J. Aerosol Sci. 18, 689e692.
www.suro, www.suro.cz/en/rms/ovzdusi/monthly-mean-values.